Analysis of the forced dynamics of a masonry facade by means of input-output techniques and a linear regression model

Structural monitoring in highly seismic areas represents a valuable source of information in both structures' behaviour understanding and identification techniques' validation. In this paper the case of a permanent monitoring system, installed in a thirteenth century monumental site, the Santa Maria of Collemaggio Basilica (Italy), is presented. Both low intensity tremors and strong motion earthquakes are recorded. Recent earthquakes provided both input (base ground accelerogram) and output (structure's response) signals. Hence a comparison has been drawn between the results from a set of Output-Only (OO) Identification techniques and a set of Input-Output (IO) algorithms. In the group of IO techniques, a rapid and easily implementable technique, based on an OLS linear regression model, is presented. Different results between the IO techniques, driven by earthquake recordings, and the OO ones, driven by ambient vibration data, emerge. Attempts are made to capture the system's dynamics, by means of a linear regression model, stressing possible research issues in the identification of non stationary systems subjected to seismic excitation. Copyright © 2019 COMPDYN Proceedings. All rights reserved.